I got this rig up and running about a week and a half ago or so, and around 00:15 - 00:45 on the 5th I noticed a severe drop in power draw and hash rate on one of my 3080ti FE cards. I was asleep during this time, and so far absolutely no amount or OC adjustments or reboots, driver updates/rollbacks, or hiveos versions have caused a positive change. It’s just running around 170W max and roughly 30MH/s (when it was previously running around the same as the rest of my 3080ti FE cards in the screengrab below). The GPU in question is GPU9 at the bottom of my flight sheet. Using the latest version of t-rex for the miner (0.25.9)
Thought it might be a good time to post and see if anyone in the community had any suggestions regarding it, or any other pointers for OC settings in general while we’re here!
I should probably mention that I am hoping to avoid flashing EVGA bios to this 3080ti as I don’t want to end up causing an issue that bricks the GPU (and have seen this as a frequent suggestion to try). Mostly I’m confused how hashrate and power draw dipped by roughly 50MH/s and 80-100W with no changes in settings/temp/anything I can see on that 3080ti (was previusly running approx 61-64 C temp wise).
Ok, some interesting notes – seems a few have quite high temps (100º+ C), BUT, someone on reddit also suggested trying lolminer v1.47 and higher, and that seems to have bumped things up a bit as well. See new screenshot below!
Always use the latest software for best results and stability, if you raise your mem clocks on the 3080ti to the highest that is stable you should get a little more performance as well.
Like I mentioned above, you should change all your clocks from negative core offsets to locked core clocks, with the goal being the lowest core clock that maintains full hashrate for each card. No need for power limits.
